Imagine a fil with no ID3 tags and filename is "ABBA - Waterlooooo.mp3"
I would like to tag the file with correct artist, title, earliest release year if a single, the single cover, the genre and style from Discogs.
It's easy to do this manually, but when you have a laaaarge collection. Not fun.
To do this manually, I right click the file, select:
Tag Source > Cover art > Discogs Artist + Title
This usually fails if artist and title is not 100% correct. In the case where it is, I sort by release year and find the best cover from. When I select the cover, I can click "Preview" which open the discogs wenpage for the release. I need that, because when I click "Ok" in mp3tag to select the release, it doesn't 'remember' anything but artist and title - I have to type in the release year and genre/style from the web page.
Is there a way to get this to work smoothly for one file assuming both artist and title is written correctly?
With the same assumption, can I make this work for a selection of multiple mp3 files?
Can I get it to find the 'closest' release to what the arist/title is - as in the case above where the title is "Waterlooooo"?
If you take the direct web source without the submenu "Cover art" you get most of the data including the cover.
You are still stuck with an album as the only grouping. And if the the album is a single, then you can get only the information for a single file.
Other programs like Musicbrainz Picard may be better to tag several files in one session.
To get basic tag data from a filename, see the documentation:
AFAIK the Discogs Scripts for Mp3tag are not designed to get metadata for Singles, but for entire Albums.
And most of the Websource Scripts for Mp3tag are not designed to work with different Singles from different artists in one run. Most of them only get metadata for one specific album from one specifiy artist (and year).
It is nearly impossible to guess if "Waterlooo" is closer to "Waterloo" than "Waterlo".
Similarity is something that we can see. Software can not "see" it the same way and has only the possibility to count the differences.
"Closer" is nothing that code can decide, only if 1 character is different, if 2 characters are different and so on.
The Websource Scripts for Mp3tag has no way to calculate something like similarity or even the Levenshtein distance.
BTW:
There is script command called $distance
But I don't see how this would help to find "similar enough" ARTIST or TITLE content from a source like Discogs or others if your current TITLE is called "Waterlooooo".
On distance and "It is nearly impossible to guess..." I was hoping for a miracle - that you guys had solved this one :- ) Guess not... But I have a ton of files where they typically lack characters like "ABBA - Waterl.mp3". That might be better?
How can I get the system to return the track "Waterlo"?
As a programmer I would search for the exact match first, then loosen up and have a more wildcard approach if nothing is returned from the query. Is that a possibility?
In theory this could work, yes.
But if you try starting with
Waterl -> no match
Waterla -> no match
Waterlb -> no match
Waterlc -> no match
...
Waterlo -> no match
...
Waterlz -> no match
Waterlaa -> no match
Waterlab -> no match
...
Waterloa -> no match
Waterlob -> no match
...
Waterloo -> finally a match
I think you see why this would take forever just for one song to get the correct spelling.
And if you include all the small differences with characters like a question mark ? or a comma , or brackets [ ] ( ), you will see why that this is really nearly impossibile.
In this specifc case there is an additional hurdle:
The Single "Waterloo" from ABBA is available in many variations.
Different languages (English, Swedish, German...), different lengths, "live"
Even if you find the correct spelling, you still need to know which version exactly matches your copy with one of the available metadata entries.
